Data export API

The data export API lets you pull your raw response data as JSON or CSV from your own scripts, a scheduled job, or a BI tool. It is read-only, server-to-server, and returns the same data as the CSV export in the admin.

The data export API requires the Pro plan. On Free and Growth you can still export everything to CSV by hand from the Responses page.

Authentication#

Every request is authenticated with your store's own API key, sent as a bearer token.

Generate your key#

In the Feedloop admin, open Settings and find the API access section. A few things to know:

• The key is shown once, at the moment you generate it. Copy it straight away, because Feedloop stores only a hash of it and cannot show it to you again.
• The key starts with fdlp_ followed by 48 hexadecimal characters.
• Generating a new key replaces the old one. The previous key stops working immediately, which is also how you revoke a key you think has leaked.
• If your store drops below the Pro plan the key stops working, and starts working again if you return to Pro. You do not need to regenerate it.

Send it on every request#

Put the key in the Authorization header as a bearer token:

Authorization: Bearer fdlp_your_key_here

Never put the key in a URL or in client-side code. It grants read access to all of your response data.

Endpoint#

GET https://feedloop.saliapps.com/api/v1/export/responses

Quick start#

curl -s https://feedloop.saliapps.com/api/v1/export/responses \
  -H "Authorization: Bearer fdlp_your_key_here"

That returns the first page of responses as JSON. The examples further down add date filtering, CSV output, and pagination.

Query parameters#

Parameter	Type	Default	Description
since	ISO 8601 date or datetime	none	Only responses submitted at or after this time, for example 2026-07-01 or 2026-07-01T00:00:00Z.
format	json or csv	json	The response encoding.
limit	integer, 1 to 250	250	Page size. Values above 250 are capped at 250.
cursor	response id	none	Return the page that follows this response id. Pass the nextCursor from the previous page.

Responses are always ordered by ascending id, which is what keeps the cursor stable as new responses arrive.

JSON response#

{
  "responses": [
    {
      "id": "cms7a5af6000nvln4ogvkij0w",
      "submittedAt": "2026-07-28T13:57:01.705Z",
      "survey": "Post-purchase NPS",
      "source": "THANK_YOU",
      "npsScore": 1,
      "orderId": "5079105710",
      "recoveryCode": "FEEDLOOP-A1B2C3",
      "answers": [
        { "question": "How likely are you to recommend us?", "type": "NPS", "value": "1" },
        { "question": "What let us down?", "type": "TEXT", "value": "Shipping was slow" }
      ]
    }
  ],
  "nextCursor": "cms7a5af6000nvln4ogvkij0w"
}

Fields#

Field	Type	Notes
id	string	The response id. This is also the pagination cursor.
submittedAt	string	ISO 8601 timestamp of the submission.
survey	string	The survey's name.
source	string	Where it was answered: THANK_YOU, ORDER_STATUS, or THEME_EMBED (the on-site pop-up).
npsScore	integer or null	0 to 10, or null if the survey had no NPS question.
orderId	string or null	The Shopify order id, or null for anonymous pop-up responses.
recoveryCode	string or null	The detractor discount code, if one was created for this response.
answers	array	One entry per answered question.
answers[].question	string	The question prompt.
answers[].type	string	NPS, SINGLE_CHOICE, MULTI_CHOICE, or TEXT.
answers[].value	string	The answer. Multiple-choice selections are joined with a comma and a space.
nextCursor	string or null	Top level. The id to pass as cursor for the next page, or null when there are no more pages.

CSV response#

Add format=csv to get a flat file instead of JSON:

curl -s "https://feedloop.saliapps.com/api/v1/export/responses?since=2026-07-01&format=csv" \
  -H "Authorization: Bearer fdlp_your_key_here" \
  -o responses.csv

The columns are:

response_id,submitted_at,survey,source,nps_score,order_id,question,answer,recovery_code

There is one row per answer, so a response with three answers spans three rows that share the same response_id. A response with no answers still emits one row, with the question and answer columns left blank, so nothing silently disappears from the file.

Pagination#

Each page returns up to limit responses (250 by default) and a nextCursor. Pass that cursor back to get the next page. When nextCursor is null, you have reached the end.

Here is a full loop in bash that walks every page, using jq to read the cursor:

KEY="fdlp_your_key_here"
BASE="https://feedloop.saliapps.com/api/v1/export/responses"
cursor=""
while true; do
  url="$BASE?limit=250"
  if [ -n "$cursor" ]; then url="$url&cursor=$cursor"; fi
  page=$(curl -s "$url" -H "Authorization: Bearer $KEY")
  echo "$page" | jq -c '.responses[]'
  cursor=$(echo "$page" | jq -r '.nextCursor // empty')
  if [ -z "$cursor" ]; then break; fi
done

Rate limits#

The endpoint allows up to 60 requests per 60 seconds per store. Beyond that it returns 429 rate_limited; wait a moment and retry. At a page size of 250 that is up to 15,000 responses a minute, far more than any single export needs.

Errors#

Errors return the matching HTTP status and a JSON body shaped like { "error": { "code", "message" } }.

Status	code	When
400	validation_error	since is not a valid ISO 8601 date.
401	auth_error	The Authorization header is missing or malformed, the key is unknown or was revoked, or the store is not on Pro.
429	rate_limited	More than 60 requests in 60 seconds.
500	internal_error	An unexpected server error.

{ "error": { "code": "auth_error", "message": "The data export API requires the Pro plan" } }